Background: Intratumoral human epidermal growth factor receptor 2 (HER2) heterogeneity of gastric cancer can be an obstacle to accurate HER2 assessment. Serum HER2, concentrations of the HER2 extracellular domain shed into the bloodstream, has a potential to compensate HER2 immunohistochemistry (IHC) but has not been scrutinized in gastric cancer. This study sought to explore the clinical utility of serum HER2 in gastric cancer. Methods: We performed a prospective multicenter trial (SHERLOCK trial) involving patients with allstage gastric or gastro-esophageal junction cancer. Serum HER2 was measured using direct chemiluminescence while tissue HER2 status was determined using IHC and fluorescent in situ hybridization. For stage IV cases, concordance between local and central laboratories in tissue HER2 assessment was also evaluated. Results: Of 224 patients enrolled, both tissue HER2 status and serum HER2 levels were successfully determined in 212 patients and 21% (45/212) were tissue HER2-positive. Serum HER2 levels, ranged from 4.5 to 148.0 ng/ml (median 10.3), correlated with tissue HER2 status (p ¼ 0.003). At a cut-off level of 28.0 ng/ml determined by receiver operating characteristics analysis, sensitivity, specificity, positive and negative predictive values of serum HER2 were 22.6%, 100%, 100% and 82.3%, respectively. All nine cases with elevated serum HER2 were tissue HER2-positive stage IV cases. Among 61 stage IV cases, the agreement rate for IHC scoring between the local and the central laboratories was 82% and tissue HER2 judgment was conflicting in five (8.2%) cases. Of these five cases, four were confirmed as false-negative and two of these four patients demonstrated elevated serum HER2. Conclusions: Serum HER2 levels correlated with tissue HER2 status in gastric cancer. Although the low sensitivity is a drawback, serum HER2 might be a useful adjunct tool to detect tissue HER2 false-negative gastric cancer.